While reading about the history of materials for jet engine shafts I came across a nice sentence that identified many common stress-raisers—features which by concentrating stress in small areas make fatigue failure more likely to occur there.
It identified “…keyways, steps, shoulders, collars, threads, holes, snap-ring grooves, and shaft surface damage.”
Many people have far more engine-build and servicing experience than I have, but I have seen these classic failures plenty of times.
Keyways are cut into ignition/alternator shaft tapers as a means of keeping the rotor properly timed to the shaft so that ignition timing doesn’t vary. Years ago I was told by old-timers that keys were unnecessary provided the rotor and shaft were lapped into “intimate contact” with abrasive powder. Just position the rotor at the desired timing and “give ‘er a smart rap with a soft hammer.” My experience has been otherwise, having seen such keyless mounting slip more than once.
But the process of cutting a keyway does concentrate stress at sharp angles, and I have seen the ends of ignition shafts break off with the failure originating at the keyway. A frequent response by manufacturers has been to just make the shaft a couple of millimeters bigger to reduce overall stress.
It was once common in the design of pressed-together roller crankshafts to give the crankpin two diameters—one for the big-end bearing’s rollers to bear upon, and slightly smaller diameters where the crankpin presses into the flywheels. This simplified assembly by allowing the flywheels to be pressed on until they made contact with the larger diameter part of the pin.
Unfortunately, unless the step where one diameter joins the other is carefully radiused and given a smooth finish, stress concentration there causes a crack to form over time, resulting in the broken crankpins I have seen. I have never, ever seen breakage in straight, stepless crankpins. I believe Harley-Davidson adopted straight pins after many years of making pins with surface features such as steps, keyways, threads, et al. Straight crankpins have no stress-concentrating surface features!
The purpose of a collar is to position either the shaft (by abutting against a bearing) or a part installed on it. They are a special case of a step, in that they are often designed with a small or nonexistent radius where the collar joins the shaft. That’s where trouble can dwell.
Many times you will see bolts or studs break at the root of the first thread. This is where the “lines of stress” that textbooks delight in talking about must funnel down from the full diameter of the fastener’s shank to the smaller thread root diameter. That concentrates the stress at that point. One common way to remove this concentration is to reduce the diameter of the fastener’s shank to slightly smaller than the thread root diameter. With my heart in my mouth, I stood at the lathe doing this to the cylinder studs of my 1970 Kawasaki H1R race engine, hoping it would work. It did—we never had another failure and Kawasaki later supplied replacement studs that looked like mine—it had reached for the same textbooks I had.
Another source of trouble in threads is surface finish (the smoother the better—cracks love to form at scratches and dings) and root radius (sharp corners must always be avoided in high-stress parts).
Finally, roll-formed threads are best at resisting fatigue because the pressure of the super-hard rolls that form them place the material’s surface in compression. Tension is required to produce cracking, but to create any tension at all, applied stress must first overcome this surface compression.
There is now a roll-forming process for internal threads as well.
The crankshaft design team knows that all its good work in avoiding sharp edges (such as providing smoothly radiused fillets where crankpins flare out at their ends to become the crank webs) can be for nothing if the holes supplying oil to the crankpins are rich in surface defects.
Wright Aero provided its large aircraft radial piston engines with multipiece forged steel crankcases but when the company tried to up-rate the 18-cylinder engines on the B-29 bomber of World War II, fatigue cracks spread from pressure equalization holes in main bearing webs during test operation at 2,600 take-off horsepower—a 25 percent increase over standard. Fatigue loves holes and discontinuities! Why do you suppose tree trunks flare gracefully at the base to form the root system? A billion years of trial and error.
Snap rings are convenient for locating gears on gearbox shafts, but the sharp corners of such grooves form a dotted line saying “fail here.” Usually, in the middle portions of gear shafts, the grooves are cut only into the projecting shaft splines and don’t give trouble. But I have seen shafts break at such grooves. As with ignition tapers, the simple treatment is to increase shaft diameter a bit.
Splines themselves, if not carefully designed, can be implemented with sharp corners. A variety of proprietary shapes exist for connecting parts with smooth, continuous shapes rather than sharp-edged ones. Google “Curvic Couplings.”
Shaft Surface Damage
When I attempted to rebuild a customer’s 50,000-mile H1 crank I nearly pinned the gauge on my hydraulic press—the force required to press mainshafts out of crank wheels was much higher than normal. Once the crank was apart I could see what had happened. All those fast miles up and down the Maine Turnpike had caused ever-so-slight relative motions between mainshafts and webs, locally creating welds that took a lot of force to break. When the pieces did separate I could see they were too damaged for reuse.
In other cranks I saw that process—called frettage—at an earlier stage, and in one case a fatigue crack had spread from the surface damage to result in a mainshaft breakage.
Why not weld the parts together as the drag racers do? It pretty much makes your crank into what the late rider/engineer Hurley Wilvert called “one-use type”—not rebuildable.
Like many others who have done this work, I’ve kept a variety of failed parts—a “museum of failure”—for the instruction and understanding they provide.